Glucocorticoid receptor-cAMP response element-binding protein ...

THE JOURNALOF BIOLOGICAL CHEMISTRY Vol. 268 No. 8 Issue of March 15 p 5353-53561993 0 1993 by The American Society for Biochemistry and Mbfechar Biology, Inc. Printed in U.S. A.

Communication Glucocorticoid Receptor-CAMP Response Element-Binding Protein Interaction and the Response of the Phosphoenolpyruvate Carboxykinase Gene to Glucocorticoids* (Received for publication, November 16, 1992) Enyu ImaiSQ,Jeffrey N. MinerT, John A. MitchellSII, Keith R. YamamotoT, and Daryl K. GrannerS** From the $Department of Molecular Physiology and Biophysics, Vanderbilt University Medical Center, Nashville, Tennessee 37232-0615 and the TDepartment of Biochemistry and Biophysics, University of California, San Francisco, California 94143-0448

The phosphoenolpyruvate carboxykinase (PEPCK) gene encodes the rate-limiting enzyme in gluconeogenesis. Glucocorticoids enhance PEPCK gene expression through a multicomponent regulatory complex. We show that a full response to glucocorticoids requires two DNA segments: 1) a glucocorticoid response unit (GRU), centered at about position -400, which contains twoaccessory factor elements (AF1 and AF2) and two glucocorticoid receptor binding sites (GR1 and GR2), and 2) a basal promoter/cyclic AMP response element (E/CRE) at about position -90, which binds the transcription factor CREB. A protein-protein interaction was observed in vitro between GR and CREB that might account for the role of the E/CRE in the glucocorticoid response of the PEPCK gene.

at which the activity of specifically bound GR depends on other nearby DNA-binding factors (6, 7). In contrast, GRmediated enhancement at “simple” GREs is conferred solely by the bound receptor. Indeed, AF1 andAF2 also serve other functions. AF1 includes a retinoic acid response element (8, 9) and an HNF-4 element’ that may be involved in tissueselective expression of hepatic genes (10); AF2 functions as an insulin responsesequence and asa phorbol ester response sequence (11,12). The interposition of the insulin and phorbol ester response sequences within the GRU may account for the inhibition of the glucocorticoid response by insulin and phorbol esters at the PEPCK promoter(2, 11-13). The responseof the PEPCKgene to cyclic AMP issimilarly complex. Control is mediated primarily through an E/CRE at -90 that is both a part of the basal promoter and a cyclic AMP response element (3, 4). The E/CRE binds the transcription factor CREB,which is phosphorylated by the cyclic AMP-dependent protein kinase inresponse to elevated cyclic AMP (14, 15); phosphorylation appears to stimulate the activity of CREB without altering its affinity for the E/CRE (16). In addition, the P3(I) element at -240 is required in certain cell contexts for a maximal response of the PEPCK promoter to cyclic AMP (17). Interestingly, P3(I) binds the transcription factor C/EBP, but not CREB, and mutationof bothP3(I)andE/CREabolishes responsiveness of the PEPCK promoter to cyclic AMP (17). Thus, a full response of the PEPCK gene to cAMP requires both the E/CRE and P3(I), with associated binding proteins. In this paper we now show that glucocorticoid responsiveness of the PEPCK gene involves a functional interaction between the GRU and the E/CRE, perhaps through a physical association of the glucocorticoid receptor with CREB. EXPERIMENTALPROCEDURES

Cell Culture and Transfection-H4IIE rat hepatoma cells were grown as described previously (18). Eighteen hours before experimentation, the cells were placed in serum-free Dulbecco’s modified Eagle’s Glucocorticoid hormones andcyclic AMP increase the rate medium. Transfection was accomplished by the calcium phosphate of transcription of the PEPCK’ gene (1,2), and eachdoes so co-precipitation method, as described previously (8, 13). Plasmids-The glucocorticoid receptor expression plasmid through a complex DNA element (3-5). For example, horwas co-transfected with various reporter plasmids where monal regulation at the GRU requires notonly the glucocor- pSVGRl indicated. The plasmid pGRUTK was created by ligating the HindIIIticoid receptor bound at two contiguous sites, GR1 andGR2, XbaI fragment of pPG44 (Ref. 19; -600 to -200 of the PEPCKgene) but also nonreceptor factors bound at two adjacent sites, AF1 into the polylinker site of the TKCAT vector. pGRETK was created and AF2 ( 5 ) . Thus, the GRU is a “composite” GRE, a locus by ligating a single copy of GRE 1.3from the MMTV promoter (5”AGCTTGTTACAAACTGTTCT-3’, 3”ACAATGTTTGACAA* This work was supported in part by Health and Human Services GATCGA-5’) into the polylinker site of the TKCAT vector. The Grants DK35107,DK20593 (to the Vanderbilt Diabetes Research series of internal deletion plasmids employed in the experiment and Training Center), andCA20535. The costs of publication of this illustrated in Fig. 3 were described by Quinn et al. (4). Radiolabeled CREB, c-Jun, and A c-Jun-[35S]Methionine-labeled article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accord- CREB, c-Jun, and Ac-Jun were produced by in vitro transcription (T7 polymerase for CREB, and SP6 polymerase for c-Jun and Acance with 18 U.S.C. Section 1734 solelyto indicate this fact. J Recipient of training support from an American Diabetes Asso- Jun; Promega) and in vitro translation in reticulocyte lysates (Promega). The CREB and c-Junexpression vectors were provided by M. ciation Mentor-based grant. (1 Recipient of postdoctoral traineeships from the American Cancer Montminy and R. Turner, respectively. Ac-Jun was prepared by digestion of the c-Jun expression vector with PstI, which cleaves the Society and the Leukemia Society. ** To whom correspondence should be addressed Dept. of Molec- DNA at a position corresponding to amino acid 222. This results in ular Physiology and Biophysics, Vanderbilt University Medical Cen- a C-terminal truncation that eliminates the bZIP (DNA binding) ter, Nashville, T N 37232-0615. Tel.:615-322-7004;Fax: 615-322-7326. region of the protein and serves as a negative control (20). Immunoprecipitation-The [35S]methionine-labeledCREB, c-Jun, The abbreviations used are: PEPCK, phosphoenolpyruvate carboxykinase; GR, glucocorticoid receptor; GRU, glucocorticoid re- or Ac-Jun proteins were incubated for 30 min at 30 “C with cytosolic sponse unit; GRE, glucocorticoid response element; E/CRE, basal extracts isolated from HeLa cells infected with either wild type promoter/cyclic AMP response element; CREB, cAMP response ele- vaccinia virus, or with a GR-expressing recombinant virus (provided ment binding protein; HRE, hormone response element; bZIP, basic zipper. * R. Hall, F. Sladek, and D. K. Granner, unpublished observation.

5353

Functional GRU and CRE Interaction

5354

by H. Stunnenberg; Ref. 21). Afterincubation, 10 volumes of HEGNO5O (10 mM Hepes, pH 8.0, 1mM EDTA, 10%glycerol, 50 mM NaCI) + 0.1% Triton X-100 was added. Themixture was then precleared by the addition of preswollen, prewashed protein A-Sepharose (100 mg/ml) to 0.03% of the final volume; the Sepharose was removed by centrifugation after 30 min a t 4 "C. GR-specific monoclonalantibody(BUGR2, provided by R. Harrison; Ref. 22) and another aliquot of protein A-Sepharose were then added, and the slurry was incubated a t 4 "C for 1 h with mixing. Sepharose was centrifuged, quickly washed four times in20 volumes of HEGN050 + 0.1% Triton X-100, transferred to fresh tubes, and washed once in HEGN050. Precipitated proteinswere eluted with sample buffer and analyzed by SDS-polyacrylamide gel electrophoresis, fluorography, and autoradiography.

construct contains the entire GRU but lacks the PEPCK basal promoter, which consists of a CAAT box, the E/CRE, and aTATA box (Ref. 4 and Figs. 2 and3). Basal and dexamethasone-induced activitywas measured in H4IIE cells cotransfected with these constructs and a GR expression vector. In contrast to pPL9, which was induced 9-foldin response to dexamethasone, pGRUTK was induced less than 2-fold (compare to pTKCAT; Fig. 2). The thymidine kinase promoter is not limiting, since a construct in which a simple GRE was ligated to pTKCAT yielded an 18-fold induction (pGRETK; Fig. 2). These results implied that some element in the PEPCK promoter between -200 and +69 is required for maximal activity of the GRU. We found that internal deletions of either the CAAT box RESULTSANDDISCUSSION and the E/CREtogether (pID6; Fig. 3) or of the E/CRE alone Glucocorticoid hormones and cyclic AMP each induce (pID3; Fig. 3) produced a marked reduction in the glucocorPEPCK mRNA in H4IIE rat hepatoma cells (Fig. 1).These ticoid response; each was stimulated approximately 3-fold by effects are additive, and atearly times aredue to anincreased dexamethasone, whereas theintact promoter (pPL9) was rate of transcription of the PEPCKgene (1,2). As mentioned stimulated 9-fold (Fig. 3). In contrast, glucocorticoid inducabove, the effect of each inducer is mediated through a com- tion was unaffected by deletion of the CAATbox alone plex DNA element. A functional interaction between these (pID11; Fig. 3) or of sequences between the E/CRE and the two complex elements, the PEPCK GRU and E/CRE, was TATA box (-75 to -40; pID7; Fig. 3). The latterregion has implied by experiments in which we compared the activity of been proposed as a glucocorticoid response element (3); howthe GRU on two different promoters. In particular, we tested ever, we failed to detect either independent GRE activity (5, the intact PEPCK promoter (from -600 to +69, in pPL9) 19) or a contribution to GRU activity by this segment (Fig. and a fusion of the PEPCK -600 to -200 region to the 3). thymidinekinasepromoter (pGRUTK; Fig. 2). The latter These results showed that maximal glucocorticoid induction of the PEPCK gene requires an intactGRU and E/CRE 1 2 3.4 and suggested the possibility that the proteins that bind to PEPCK -I these elements might interact. The factors that associate with AF1 and AF2 have not yetbeen identified, but it is clear that the receptor binds to two sites in the GRU (GR1 and GR2) 7 and thatCREB bindsto the E/CRE (4,5,17). Towhether test e the receptor andCREBinteract, [35S]methionine-labeled Fold Induction 1.0 5.0 4.3 9.3 FIG. l. Dexamethasone and CAMP induce accumulation of CREB was mixed with extracts of HeLa cells infected with a PEPCK mRNA.H4IIE rathepatoma cells were incubated in serum- GR-expressing recombinant vaccinia virus, or with wild type free Dulbecco's modified Eagle's medium containing noadditive (lane vaccinia virus. After incubation, the receptor and associated I), 0.1 mM 8-(4-~hlorophenylthio)cAMP(lane 2 ) , 0.5 p~ dexamethasone (lane 3 ) ,or both (Lune 4 ) for 6 h. Total RNA was isolated and quantitated using a primer extension assay. Data from this representative experiment are expressed as the -fold increase of PEPCK mRNA in treated cells compared to cells that received no treatment (control). Differences in the amountof total RNA added to theprimer extension assay are accounted for by expressing the ratio of PEPCK mRNA to that of calmodulin mRNA, which is not affected by either of these inducers. The ratios of PEPCK + calmodulin mRNA were 0.4, 2.0, 1.7, and 3.7 for lanes 1-4, respectively.

c.QmMu

-600

PR9 plD6

plD3 plDl1 plD7

-5W

-400

-3W

,200

Ld.4 .lo0

140121

3.6r08

GI1

*

nf"

R

:

&a-

Blul

1WfO

896tl3l

52.6

e~

E

l78t38

CAAT

CRE

9.0r1.3

2.8tO.2

57110 59118

132a7

48Ofl52

9.3~2.8

1083t112

8.4213

T.\rA

FIG. 3. The PEPCK proximal promoter region is involved in glucocorticoid regulation through the GRU. The various internal deletion mutations of PEPCK proximal promoter region PPL9 90213 analyzed are illustrated a t left. The components of the glucocorticoid PGRUTK 19fO.J response unit (GRU) and thebasal promoter of the PEPCKgene are illustrated as in Fig. 2. The boundaries of the deletions in the various constructs are as follows: pID6, -129/-86; pID3, -95/-86; pID11, l8OiO9 -log/-101; pID7, -75/-40. Ten pg of the individual constructs were cotransfected into H41IE cells with 5 pgof glucocorticoid receptor .Of0.l pTK.CAT expression vector (pSVRG1) and 2.5pgof aninternal control FIG. 2. The GRU functions poorly through the thymidine (PRSVL-AD~'),which was used to monitor transfection efficiency. kinase promoter. Ten pg of the plasmids illustrated and 5 pg of the The CaPO., co-precipitation method was used for transfection. After glucocorticoid receptor expression plasmid pSVGR1 were cotrans- transfection the H4IIEcells were placed in serum-free medium in the fected intoH4IIE cells by the calcium-phosphatemethod. After presence or absence of 0.5 p~ dexamethasone for 24 h. Basal level transfection the cells were placed in serum-free medium in the pres- expression of CAT activity of the various constructs, corrected for transfection efficiency, was compared to thatof the wild type vector, ence or absence of 0.5p~ dexamethasone for 18 h. The -fold induction of chloramphenicol acetyltransferase activity is expressed as mean which was assigned a value of 100% in each experiment. The basal S.E. for a minimum of five independent transfections. The construc- promoter elements were mapped in experiments using CV-1 cells (41, tion of the plasmids pGRUTK and pGRETK is described under whereas all the experiments reported in this paper were done using "Experimental Procedures." The glucocorticoid response unit is lo- H4IIE rat hepatoma cells. This may account for the difference in cated from base pairs -455 to -349 and includes two glucocorticoid basal activity notedin the two experiments.The ratioof CAT activity AF1 ( ), and AF2 ( 4). The basal pro- in dexamethasone induced versus control cells is expressed as the receptor binding sites moter region of PEPCK gene includes a CAAT box ( :), E/CRE ( ), mean & S.E. of four experiments for the four members of the ID series and seven experiments for pPL9. and a TATA box ( ).

* 1%

e),

CI,

1

Functional GRU and CRE Interaction

5355

proteins were immunoprecipitated with a monoclonal anti- thyroid family of hormones, and a nonreceptor, intracellular body directed against GR (22). The immunoprecipitateswere protein, such as CREB, in thecase of hormones thatbind to subjected to SDS-polyacrylamidegel electrophoresis and au- receptors located in the cell surface. Although such simple toradiography. As shown in Fig. 4, labeled CREB was precip- response elements can autonomously enhance transcription itated in the presence (lune 9)but not in the absenceof GR upon association of their cognate factors, they lack the ver(lane 10). As expected from previous work (23), co-precipita- satility of more complex regulatory regions whose behavior is tion of a GR.c-Jun complex was also observed (lanes 7 and defined by the interactions of two or more distinct protein 8 ) , whereas labeled AcJun (a C-terminal truncation of cJun factors (6, 7). Many types of physiologic regulation require that lacks the DNA-binding region) was unable to form a counter-balancing positive and negative control mechanisms, may necessitate that either the positive complex with GR (lanes 11 and 12). Although quantitation of and certain situations or negative activity be dominant. In addition, the concerted such assays is difficult, it appears that the GR- complex CREB may form moreefficiently than does the GR. Jun complex. In actions of several different hormones ona gene may result in additive or synergistic effects. contrast to earlier studies (23), these reactions did involve not Composite HREs,which contain sequences recognized both chemical cross-linking agents. by receptors and by non-receptor factors, appear capable of Because both the receptor and CREB are DNA-binding proteins, it was conceivable that their co-immunoprecipita- just this sortof regulation (6,7). Forexample, communication tion resulted from association of the proteins with contami- between the transcription factor AP1 and various members nating DNA in the extracts rather than with each other. Laiof thesteroid/thyroidhormonereceptor family has been and Herr (24) demonstrated that such protein-DNA interac- described in the regulation of the proliferin and collagenase tions could beeliminated by the additionof ethidium bromide, genes by glucocorticoids (23, 26-28), the stromelysin gene by a DNA-intercalatingagent.Unfortunately, we found that retinoic acid (29), the ovalbumin gene by estrogens (30), and ethidium bromide inhibited the interaction of our monoclonal the osteocalcin gene by vitamin D and retinoic acid (31-33). antibody with the receptor, making the ethidium bromide test Direct interaction of AP1 with particular receptors hasbeen ineffective for our purposes. However, severaladditional lines demonstrated in some of these cases and implied in others. One composite element, plfG, a 25-base pair sequence from of argument suggest strongly that the receptor and CREB interact directly. First, depletionof contaminating DNA with the proliferin gene promoter, is recognized both by GR and micrococcal nuclease or DNase I had no effect on the effi- by AP1, a transcription factorcomprised of c-Jun homodimers ciency of coprecipitation. Second, the amounts of coprecipi- or c-Jun/c-Fos heterodimers. The ratio of c-Fos and c-dun tated proteinwere insensitive to the addition of high levels of activities determines whether glucocorticoids have no effect, nonspecific DNA. Third, the reactionswere carried out with a positive effect, or a negative effect on the proliferin gene. of AP1, dexamethasone has noeffect on low concentrations of CREB and receptor in the presence of Thus, in the absence a vast excess of other proteins; the artifact pointed byout Lai plfG; however, the hormone stimulates transcription when represses transcription and Herr (24)was observed only a t very high concentrations c-Jun is in excess relativeto c-Fos and whenc-Fosisin excess of c-Jun. Aphysicalassociation of the “interacting” proteins. Minimal DNA sequences that constitute functional HREs between GR and Jun, through this assembly of contiguous have been defined for several hormones (25). These generally elements, appears to be involved in composite regulation of are less than 20 base pairs in length and aredefined by their the proliferin gene (23). The GRU in the PEPCK gene is itself a composite element abilitytobind a factor that regulatestranscription.This factor is the intracellular receptor in the case of the steroid/ in which factors bound a t two accessory factor sites,AF1 and of GR bound two to immediately AF2, are essential for activity (5). We now show that a full response to adjacent sites IMMUNOPPT INPUT glucocorticoids requires both the GRU and E/CRE in a func+ + + GR + - + - + tional interaction that is accomplishedovera distance of cJun + + - - - + + “ ” about 300 base pairs. That is, the activity of the composite - -++- CREB - - + + - GRE itself can be further modulated by physical association ” “ + + AcJun - - - - + + withadditionalfactorsbound a t remotesites. It will be -68 interesting in future work to determine whether these inter-43 actions, as well as others that havebeen inferred or detected 29 a t loci that contain DNA binding sequences for only one of the putative interacting factors (see Refs. 6,7, 23, and 26-33) confer their effects via a single molecular mechanism or via multiple distinct pathways. 1 2 3 4 5 6 7 8 9 101112 Our finding of an interaction between the glucocorticoid FIG.4. GR and CREB interact in vitro. Radiolabeled CREB, cJun, and AcJun wereincubated in extractswith or without unlabeled receptor and CREB may be considered unsurprising, given GR, as described under “Experimental Procedures.” An aliquot was the known structural relationship of CREB and Jun (15), and removed and analyzed by SDS-polyacrylamide gel electrophoresis it is possible that other membersof the superfamily of bZIP (INPUT, lanes 1-6). The remainder of each reaction was subject to proteins form similarinteractions. Nevertheless, it is also immunoprecipitation with an anti-GR monoclonal antibody (22), as described under “Experimental Procedures.” Proteins were eluted clear that GRdoes not simplyassociate with all bZIP proteins from the protein A-Sepharose with sample buffer and analyzed by (23).3 The detailed mechanism of the interaction between SDS-polyacrylamide gel electrophoresis, fluorography, and autoradi- these members of two different families of transcription facography. The results of this procedure are shown in lanes 7-12 under tors remains tobe investigated. ZMMUNOPPT (immunoprecipitation). The relative migration of Cyclic AMP andglucocorticoids influencemany of the same expected from their CREB is slightly slower thatthatofJun,as physiologic processes. In some cases one or the otherof these molecular mass (43 and 36 kDa, respectively). The migration of standard proteins in this system(bovine serum albumin, 68 kDa; effectors plays a permissive role, and numerous additive and ovalbumin, 43 kDa; carbonic anhydrase, 29 kDa) is shown at right. synergistic responses have been described (34). A functional

.

Ac-Jun, a truncated form of c-Jun, migrates much more rapidly than either of the other proteins.

M. Vivanco and K. R. Yamamoto, unpublished observation.

Functional GRU and CRE Interaction

5356

relationship between the glucocorticoid and CAMP response elements, through interaction of the proteins that bind to them, may be one way that these physiologic effects are achieved. Acknowledgments-We thank the colleagues cited in text for providing materials used in this study, T. Weil and R. O'Brien for a critical review of the manuscript, R. Printz for help in making the figures, and D. Caplenor for preparation of the manuscript. REFERENCES 1. Lamers, W. H., Hanson, R. W., and Meisner, H. M. (1982) Proc. Natl. Acad. Sci. U. S. A . 79,5137-5141 2. Sasaki, K., Cripe, T. P., Koch, S. R., Andreone, T. L., Petersen, D. D., Beale, E. G., and Granner, D. K. (1984) J. Biol. Chem. 259,15424-15251 3. Short. J. M.. Wvnshaw-Boris. A,. Short. H. P.. and Hanson. R. W. (1986) . . J. Biol. Chemr 261,9721-9726 4. Quinn, P. G., Wong, T. W., Magnuson,, M. A,, Shabb, J. B., and Granner, D. K. (1988) Mol. Cell. Biol. 8. 3467-3475 5. Imai, E.,'Stromstedt,P.-E., QGnn,, P . G., Carlstedt-Duke, J., Gustafsson, J.-A., and Granner, D. K. (1990) Mol. Cell. Biol. 10,4712-4719 6. Miner, J. N., and Yamamoto, K. R. (1991) Trends Biochern. Sci. 16,423426 7. Lucas, P. C., and Granner, D. K. (1992) Annu. Reu. Biochem., 6 1 , 11311173 8. Lucas, P. C., O'Brien, R. M., Mitchell, J. A., Davis, C. M., Imai, E., Forman, B. M., Samuels, H. H., and Granner, D. K. (1991) Proc. Natl. Acad. Sci. U. S. A. 88. 2184-2188 9. Lucas, P. C., Forman, B. M., Samuels, H. H., and Granner, D. K. (1991) Mol. Cell. Biol. 1 1 , 5164-5170 10. Sladek, F. M., Zhong, W., Lai, E., and Darnell, J. E., Jr. (1990) Genes & Deu. 4,2353-2365 11. O'Brien, R. M., Lucas, P. C., Forest, C. D., Magnuson, M. A,, and Granner, D. K. (1990) Science 249,533-537 12. O'Brien, R. M., Bonovich, M. T., Forest, C. D., and Granner, D. K.. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,6580-6584

13. Magnuson, M. A., Quinn, P. G., and Granner, D. K. (1987) J. Biol. Chem. 262,14917-14920 14. Hoeffler, J. P., Meyer, T. E., Yun, Y., Jameson, J. L., and Habener, J. F. (1988) Science 2 4 2 , 1430-1433 15. Gonzalez, G. A., Yamamoto, K. K., Fischer W. H., Karr, D., Menzel, P., Biggs, W., 111, Vale, W. V., and Montmihy, M. R (1989) Nature 3 3 7 ,

__

749-769 . ."

16. Quinn, P. G., and Granner, D. K. (1990) Mol. Cell. Biol. 10,3357-3364 17. Liu, J., Park, E. A., Gurney, A. L., Roesler, W. J., and Hanson, R.W. (1991) J. Biol. Chem. 2 6 6 , 19095-19102 18. Forest, C. F., O'Brien, R. M., Lucas, P. C., Magnuson, M. A,, and Granner, D. K. (1990) Mol. Endocrinol. 249,1302-1310 19. Petersen D.D., Magnuson, M. A,, and Granner, D.K. (1988) Mol. Cell. Biol. &96-104 20. Miner, J. N., and Yamamoto, K. R. (1992) Genes & Deu. 6,2491-2501 21. Schmid W. Striihle U. Schiitz, G., Schmitt, J., andStunnenberg, H. (198d) E d B O J. 8:226772263 22. Gametchu, B., and Harrison, R. W. (1984) Endocrimlog 114,274-279 23. Diamond, M.I., Miner, J. N., Yoshinaga, S. K., and $amamoto, K. R. (1990) Science 249,1266-1272 24. Lai, J.-S., and Herr, W. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 6958COC9

25. BerGrM. (1989) Cell 56,335-344 26. Jonat, C Rahmsdorf H. J., Park K.-K., Cato, A.C. B., Gebel, S., Ponta, H., and Herrlich, P: (1990) Cell'62, 1189-1204 27. Schule, R., Ran arajan P I Kliewer, S., Ransone, L. J., Bolado, J., Yang, N., Verma, I. and Eians,R. M. (1990) Cell 62,1217-1226 28. Yang-Yen H.-F., Chambard, J.-C., Sun, Y.-L., Smeal, T., Schmidt, T. J., Drouin 'J. and Karin M. (1990) Cell 62,1205-1215 29. Nicholsoi R. C Made; S Naal, S., Leid, M., Rochette-Egly, C., and Chambdn, P. (1990) EMBO J.%, 4443-4454 30. Gaub, M.-P., Bellard, M., Scheuer, I., Chambon, P., and Sassone-Corsi, P. (1990) Cell 63,1267-1276 31. Schule R. Umesono K. Mangelsdorf D. J., Bolado, J., Pike, J. W., and Eva& R. M. (199d) Clll 61,497-502 32. Owen T:A., Bortell R., Yocum, S. A. Smock, S. L., Zhang, M. Abate C., Shhhouh V., Arinin N., Wri ht, k L , Van Wunen, A. J., Stein, J: L.,, Curran, 'f., Lian, J. b., and itein, G.'S. (1990) Proc. Natl. Acad. Scr. U. S. A . 87,9990-9994 33. Ozono, K., Liao, J., Kerner, S. A., Scott, R. A,, and Pike, J. W. (1990) J. Biol. Chem. 2 6 5 , 21881-21888 34. Granner, D. K. (1979) in GlucocorticoidHormone Action (Baxter, J. D., and Rousseau, G., eds) Vol. 12, pp. 593-609, Springer-Verlag, Heidelberg

h.,